Polyhydroxyl compounds have become of great commercial importance in various fields. They are used on a large industrial scale, for example, for the manufacture of non-ionic surface active compounds, as antifreezes, as moisturizers and plasticizers and as starting components for the production of synthetic resins such as polyester and polyether resins. Polyhydric alcohols are at present obtained from naturally occurring substances such as sugar or cellulose materials or synthesized by the oxidation of petroleum derivatives.
In view of the world food situation, it is undesirable to use naturally occurring substances as raw materials for industrial products if these substances can be used as carbohydrate sources for nutrition. On the other hand, in view of the shortage of petroleum sources, the price of products which are dependent upon petroleum is constantly increasing. Moreover, the supply of petroleum products is not ensured in the long term.
It would therefore be desirable to find processes for the manufacture of polyhydroxyl compounds from raw materials which are independent of petroleum and other naturally occurring substances.
Since the work of Butlerow and Loew (Ann. 120, 295 (1861) and J. pr. Chem., 33, 321 (1886)) in the last century it has been known that hydroxyaldehydes and hydroxy ketones are formed from the condensation of formaldehyde hydrate (hereinafter the term "condensation of formaldehyde" should always be understood to mean "condensation of formaldehyde hydrate with itself") under the influence of basic compounds such as calcium hydroxide or lead hydroxide. Since formaldehyde can be obtained from coal or natural gas by way of methanol, this would in theory be a possible source of hydroxyl compounds which would be independent of petroleum. Polyhydric alcohols could then be synthesized by electrolytic reduction or by catalytic or chemical hydrogenation.
However, in spite of several proposals for the synthesis of polyhydroxyl compounds by the condensation of formaldehyde, no commercially viable process has yet been developed for such a synthesis, since no one has succeeded in synthesizing mixtures of polyhydroxyl compounds in which the hydroxyl functionality is clearly reproducible. Moreover, the known processes give rise to hydroxyaldehyde and hydroxyketone mixtures which are difficult to hydrogenate and only with very large quantities of catalyst. This high catalyst consumption has hitherto indicated that the synthesis of polyhydroxyl compounds by the autocondensation of formaldehyde hydrate is uneconomic. This has prevented the condensation of formaldehyde hydrate from being used as a basis for a commercial process for the synthesis of polyhydric alcohols.
Due to the disproportionating reaction of formaldehyde to methanol and formic acid which takes place at the same time, only moderate yields have been obtained by the known processes so that the working up of the aqueous or aqueous/alcoholic solutions obtained involved considerable costs which rendered the process uneconomical.
It is known that the disproportionation of formaldehyde into methanol and formic acid is very powerfully catalyzed by basic compounds. It has been shown by Pfeil, Chemische Berichte 84, 229 (1951) that the reaction velocity of this co-called Cannizzaro reaction depends upon the square of the formaldehyde concentration whereas the reaction velocity of formaldehyde polyaddition (C--C linkage) depends directly upon the formaldehyde concentration (Pfeil and Schroth, Chemische Berichte 85, 303 (1952)). The proportion of the desired polyhydroxyl compounds obtained to the quantity of methanol and formic acid produced is therefore shifted in favor of the unwanted compounds as the aldehyde concentration increases. In most of the known processes, it is therefore proposed that the condensation of formaldehyde to hydroxyaldehyde and hydroxyketones should be carried out in solutions having a low formaldehyde concentration in order to keep the quantity of by-products as low as possible. In that case, the water used as solvent must subsequently be removed by distillation to recover the hydroxyaldehydes and hydroxyketones formed in the process. This involves considerable energy costs due to the high heat of evaporation of water. Processes for the condensation of formaldehyde from dilute aqueous solutions are therefore uneconomical. Moreover, if distillation is prolonged, the hydroxyaldehydes and hydroxyketones undergo considerable decomposition and discoloration reactions.
A process for the preparation of aliphatic hydroxyaldehydes in which a 40% formalin solution is reacted with thallium or thallium hydroxide has been described in German Patent No. 822,385. However, this process is of doubtful value in view of the toxicity of thallium. Moreover, thallium hydroxide is difficult to obtain. The yields of this process are relatively low, ranging from 70 to 80%.
With a view to preventing the Cannizzaro reaction, it has also been proposed to react formaldehyde solutions with calcium hydroxide or lead hydroxide in the presence of methanol, ethanol or other polar organic solvents as described in German Patent No. 830,951 and Gorr and Wagner, Biochemische Zeitschrift, 262, 361 (1933). However, the addition of organic solvents again reduces the formaldehyde content of the solution. These processes would therefore also seem to be uneconomical in view of the additional energy costs required for evaporating the added solvent to work up the hydroxyaldehydes and hydroxyketones formed.
A process for the preparation of oxy-oxo compounds in which aqueous formaldehyde solutions at concentrations of up to 30% are reacted with lead oxide or lead acetate and inorganic bases to form sugar-like compounds which reduce Fehling's solution in the cold has been described in German Patent No. 884,794. In this process, however, the formaldehyde solution must be heated for 7 to 8 hours; the volume/time yield is therefore unsatisfactory. The relatively low yields (approximately 80%, based on the quantity of formaldehyde put into the process) are also by no means satisfactory.
A process for the preparation of hydroxyaldehydes and hydroxyketones in which the exothermic condensation of formaldehyde with itself is regulated by the controlled addition of inorganic or organic bases to a formaldehyde solution containing lead, tin, calcium, barium, magnesium, cerium or thorium compounds and a compound which is capable of enediol formation, such as glucose, ascorbic acid, fructose, benzoin, glycol aldehyde, erythrose, reductose, invert sugar or condensation products of formaldehyde, has been disclosed in U.S. Pat. No. 2,224,910. Although this process gives rise to a mixture of hydroxyaldehydes and hydroxyketones from relatively concentrated formaldehyde solutions without the addition of organic solvents, this advantage is offset by various disadvantages. If the reaction is carried out at a low pH value, the reaction product consists mainly of hydroxyaldehyde and hydroxyketone mixtures having a low hydroxyl functionality. Moreover, only moderate reaction velocities are attained at low pH values, so that the volume/time yields obtained in this variation of the process are not satisfactory. To overcome these disadvantages, it is recommended in the Specification to start formaldehyde condensation at a low pH value and to complete it at a higher value. However, at pH values of 7 or higher, lead catalyzed formaldehyde condensation proceeds so rapidly, spontaneously and uncontrollably, that it is not possible by this variation of the process to obtain mixtures of hydroxyaldehydes and hydroxyketones with a reproducible distribution of components. The reaction times and conditions can no longer be accurately controlled. Furthermore, it is well known that in an alkaline medium and at elevated temperatures, hydroxyaldehydes, hydroxyketones and monosaccharides decompose into dark colored compounds in part containing carboxyl groups.
A major disadvantage of the processes previously known is that the substances used as the source of formaldehyde are aqueous formalin solutions. As is well known, these are obtained on an industrial scale by a multistage process of absorption of formaldehyde from formaldehyde-containing synthesis gases in water in a series of absorption columns, followed by removal of the water by distillation to concentrate the product. These steps of the process which have up to now been necessary render the manufacture of sugar aldehydes and ketones from formaldehyde (hereinafter referred to as "formoses") and of the sugar alcohols (hereinafter-referred to as "formite") obtained from them by hydrogenation a relatively uneconomical process. It is therefore an object of the present invention to provide an economical, variable, reproducible and commercially generally applicable process by which formose-sugar mixtures of various desired molecular compositions can be produced in high yields. This process should be capable of providing colorless formoses and formites and, if desired for special purposes, it should also be able to give rise to sugar mixtures with a reddish yellow color which are already strongly caramelized and which may then be used for the applications mentioned above.